Positions

Company information

Enefit / Eesti Energia is a company which operates in the Baltic Sea electricity and gas markets and in the international fuel market. We have the most diverse energy portfolio in the Baltic Sea region: we produce energy from oil shale, biomass, tire chips, municipal waste, wind, sun, and water. We use oil shale to produce liquid fuels – shale oil and oil shale gasoline as well as electricity and heat.

Website

Introductory paragraph

We have a dedicated team that promotes and leads the innovations; we name this team E-Lab. The team consists of highly skilled specialists in the field of software engineering, data science, and cloud engineering. We test different solutions in cooperation with other units and external partners in the framework of short-term prototyping projects. Data Scientists participate in E-Lab as core team members, providing their expertise and knowledge on machine learning, data analysis, and optimization.

The Machine Learning engineer would work closely with data scientists in order to deploy their model. They would also work closely with software engineers who deal with back-end and infrastructure.

ML engineer would work starting from the feature engineering phase, creating pipelines for data cleaning, transforming and then training and predicting for the ML model in production.

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: There are offices in Tallinn and Tartu. Part of the team is in Tallinn, another part in Tartu.

Format: Hybrid. Physical presence is asked on Tuesdays (if no lesson at university), one week in Tallinn, the other in Tartu. Transport is paid by the company. Other days can be whether remote, whether at the office.

Language: English.

Expected general assignments

• Share your code
• Present your work.
• Integrate MachineLearning models.
• Use MLOps tools
  

Expected topics for master’s thesis

(Technical) expectations for applying student

• Good at machine learning
• MLOps
• Strong IT skills
• Presentation skills
• Interest for energy industry

Supervisor designated by the partner

Why should you join us?

AI is increasingly used in the electricity sector.

Strong team with several data scientists and developers, so we learn a lot from each other.

Application process

In addition to your CV, answer the following question:

• Why would you be a good fit for this position?

2nd round:

• work assignment
• an interview with the supervisor

Company information

Enefit / Eesti Energia is a company which operates in the Baltic Sea electricity and gas markets and in the international fuel market. We have the most diverse energy portfolio in the Baltic Sea region: we produce energy from oil shale, biomass, tire chips, municipal waste, wind, sun, and water. We use oil shale to produce liquid fuels – shale oil and oil shale gasoline as well as electricity and heat.

Website

Introductory paragraph

We have a dedicated team that promotes and leads the innovations; we name this team E-Lab. The team consists of highly skilled specialists in the field of software engineering, data science, and cloud engineering. We test different solutions in cooperation with other units and external partners in the framework of short-term prototyping projects. Data Scientists participate in E-Lab as core team members, providing their expertise and knowledge on machine learning, data analysis, and optimization.

Data scientists work on diverse themes: renewable electricity production forecast, consumption forecast, electricity price forecast, data analysis for electricity production, for electric vehicles, energy storage, etc.

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: There are offices in Tallinn and Tartu. Part of the team is in Tallinn, another part in Tartu.

Format: Hybrid. Physical presence is asked on Tuesdays (if no lesson at university), one week in Tallinn, the other in Tartu. Transport is paid by the company. Other days can be whether remote, whether at the office.

Language: English.

Expected general assignments

• Create features to make better predictions.
• Deliver data analysis.
• Share your code
• Present your work.
• Finally, integrate your model with developers.
  

Expected topics for master’s thesis

• Probabilistic forecasting of electricity prices
• Reinforcement Learning for battery scheduling

(Technical) expectations for applying student

• Good at mathematics, especially probability
• Deep learning
• Presentation skills
• Interest for energy industry

Supervisor designated by the partner

Why should you join us?

Interesting topics. You can apply your skills to an industry that matters: energy. Due to more & more renewable energy, electricity production is less known and requires AI to make the system stable.

Strong team with several data scientists, so we learn a lot from each other.

Application process

In addition to your CV, answer the following questions:

• What kind of model would you try for time-series forecasting? For example, predicting electricity prices for the next 12 hours.
• Why would you be a good fit for this position?

2nd round:

• work assignment
• an interview with the supervisor

Company information

GPV Estonia (GPV Group A/S)

Website

Introductory paragraph

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Elva, Valga mnt 7a.

Format: Hybrid. Periodic presence in production is required for testing, measurements, piloting activities.

Language: Estonian and/or English.

Expected general assignments

• Map current inter-process flows (production unit → warehouse → another production unit), including routes, frequencies, and loads.
• Develop a data-driven business case (baseline vs AMR/AGV, including ROI and payback time).
• Model traffic and charging strategies (single robot vs small fleet).
• Propose suitable technological solutions (manufacturers, models), including risks (e.g., MiR / Omron / OTTO / Geek+ with top modules).
• Design and execute a limited-scope pilot in production; prepare implementation recommendations and documentation.
  

Expected topics for master’s thesis

• Robotics and process automation in production intralogistics
• Application of machine vision systems for AMR navigation and obstacle detection
• Impact of AMRs on lead times, labor costs, and process efficiency across multiple production units
• Role of fleet management systems in coordinating and optimizing AMRs
• Design of safe movement, traffic flows, and zone management in compliance with ISO 3691-4
• Integration of AMR solutions with existing production and warehouse systems (e.g., ERP/MES)
• Data-driven optimization and simulation models for improving intralogistics processes
• Human–robot interaction (HRI) in production environments, including safety and user experience

(Technical) expectations for applying student

• Interest in robotics, autonomous systems, and intralogistics automation
• Basic knowledge of programming (e.g., Python) and/or data analysis
• Previous experience with robotics (e.g., SLAM, ROS) or industrial automation is a plus
• Strong analytical and problem-solving skills; ability to work in a structured manner
• Good written and verbal communication skills (including technical documentation)
• Willingness to conduct measurements, tests, and pilot activities in a real production environment
• Independence and initiative, with the ability to collaborate across teams (engineering, production, IT)

Supervisor designated by the partner

Kristjan Piir, Engineering Manager

Severin Israel, Senior Operations Manager

Jaan Oder, IT Manager

Why should you join us?

Opportunity to participate in a real factory pilot project: you will perform data-driven analysis, test AMR solutions (including top modules), and see the direct impact of your work on production processes.

You will gain hands-on experience in:

• Simulation
• Process analysis and optimization
• Safety standards (ISO 3691-4)
• Implementation of smart intralogistics solutions

Application process

In addition to your CV, answer the following question:

• What motivates you to work with AMR applications in production intralogistics? What do you want to learn from this project?
• Describe a specific example (project, studies, or work) where you analyzed a process or optimized a workflow. What approach and tools did you use, and what was the result?
• What practical experience do you have with robotics, automation, or data analysis (e.g., Python, simulation, ROS, sensors)? Provide concrete examples.
• Imagine AMRs are used between production units. What key challenges (e.g., safety, traffic, integration, efficiency) do you foresee, and how would you address them?

2nd round: an (online) interview, where your motivation, technical skills, and fit with the team is discussed

Company information

HiveNavigator

Website

Introductory paragraph

HiveNavigator develops beekeeping technology where we turn the beehive from a “black box” into a measurable and predictable system. We build big-data-driven remote monitoring and decision support for bee colonies, combining sensor data (audio, temperature, RH, etc.) with machine learning to enable predictive, needs-based beekeeping. This is the development of sector-changing technology.

The aim of the collaboration is to develop machine learning models based on beehive audio and sensor data for the early detection of beekeeping-related events.

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Remote

Format: Hybrid; at the beginning we prefer regular in-person meetings in Tartu/Viljandi (e.g. 1–2 days per week).

Language: Estonian or English.

Start of the collaboration: if possible, in April 2026

Expected general assignments

• Preprocessing and quality control of hive audio data
• Feature engineering for audio time-series (spectra, patterns, event candidates)
• Synchronization of sensor data (temperature, RH) with audio data
• Linking beekeeper-logged events and behaviours to the data based on timestamps
• Integrating event labels (ground truth) into the dataset
• Developing machine learning models for beekeeping events (for example):
  • early swarm risk model
  • queen presence/absence/acceptance
• Model evaluation (false alarm rate, precision/recall, generalization across different hives)
• Visualization of results and a simple prototype scoring pipeline
• Collaboration with hardware and beekeeping teams for event interpretation

Expected topics for master’s thesis

• Noise-robust learning in noisy biological time-series signals
• Use of autoregressive models to predict sequences of micro-events
• Incorporating time-synchronized multisensory data and human observation data (event logs) into models
• Modelling early-warning systems on weakly labelled datasets
• Generalization of machine learning models across different colonies

(Technical) expectations for applying student

• Interest in working with noisy real-world datasets
• Ability to solve problems independently
• Strong Python skills
• Basic knowledge of machine learning and neural networks
• preferably experience with time-series or audio signal processing

Supervisor designated by the partner

Why should you join us?

Application process

In addition to your CV, answer the following question:

1. Describe a project, course assignment, or personal work where you handled complex, real-world, or imperfect data. What was the problem and how did you approach it?
2. How would you approach a situation where event labels are inaccurate or partially missing?
3. Which models would you consider for time-series prediction, and why?

2nd round will comprise of a technical interview and a short mini-case discussion based on a dataset.

Company information

IndiGO Europe

https://indigotech.com/

https://clevon.com/about/

Introductory paragraph

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tartu.

Format: Hybrid: office teamwork + individual remote work.

Language: English.

Expected general assignments

Among the numerous challenges for automated vehicles (AV), indiGO’s interest for the DIGILINK program fall into one of two:

A) Engineering solutions for (semi-)automatic creation of HDMaps: HDMaps are information dense data structures which support the AV by providing info such as:

• cm-accurate street & lane geometry
• traffic sign locations
• pedestrian crossings
• regulatory information about right-of-way; to name a few.

While visual editors assist in manual creation of HDMaps, the work is very laborious and scales poorly. On the other hand, composing information from: driven logs (e.g. GPS traces), satellite/aerophoto imagery, and other maps (e.g. OpenStreetMap) has potential to automate the creation of HDMaps.

B) Automated Driving robot’s Operational design domain (ODD) detection and management: an AV functional capabilities are limited, it should only operate within the scope of those limits. For instance, unexpected road works or extreme weather conditions may fall out of the ODD of the vehicle, in which case the system should be able to detect this automatically and respond to the situation accordingly (e.g. ask for a teleoperator to take over driving and return the vehicle into a state where it can resume automated driving).

Expected topics for master’s thesis

General assignments for topic A:

• Research and prototype methods for (partially) automatic creation of Lanelet2-format HDMaps
• Work with different data sources, including open datasets, sensor logs, public geo-information
• Helping design, build a pipeline that fuses the information sources and applies the processing; helping set up the needed software services/infrastructure
• Ensuring the solution can scale to different regions (mainly Europe and US)
• Implementing validation solutions (e.g. automated tests) to ensure the automatically produced HDMaps works with indiGO-s AV

General assignments for topic B:

• Systematic assessment for matching different sensory data sources & ML models to suitable ODD detection sub-criteria
• In-depth prototyping and validating of one or several ML model’s capabilities to detect certain ODD criteria– for example: traffic density; road grip conditions; traffic sign detection illumination conditions (exact criteria to be agreed upon)
• Fine-tuning and training of ML models
• Working with sensor logs (including video) to test the detection module(s)
• Helping design & implement a framework/software module that can combine different ODD criteria detectors ensure its’ readiness for real-time performance constraints during execution.

(Technical) expectations for applying student

• Comfortable working with diverse datasets and experimenting with different data processing techniques
• Experience with Python + Pandas or R
• Knowledge of Linear Algebra and statistics
• Experience working with geo-info, sensor data and other robotics-related context
• Some previous experience in computer-vision related topics
• Experience with C++ is a big bonus
• Driven to allocate individual effort for search literature for finding existing similar works in a systematic fashion
• Capacity working within an English-language company; not afraid to communicate mistakes, blockers and express knowledge gaps (in other words ask stupid questions)

Supervisor designated by the partner

Why should you join us?

• You’ll work hands-on with state-of-the-art hardware and sensor data, bringing autonomous vehicles from our lab to the streets of Tartu (and eventually the world).
• AVs are the future of transportation, and you’ll be building technology that will shape how cities move.
• You’ll literally be building a self-driving vehicle. Something very few people can put on their CV.
• You’ll be working with the top specialists in the field.

Application process

In addition to your CV, answer the following questions:

• Briefly summarize the one software/IT-related project (a course project or hobby) that you are most proud of, and why. What was your role in that project?
• Tell us about your existing experience and technical skills relating to robotics (or other similar domains)
• Tell us briefly about a project where you had to apply linear algebra, probability theory or statistics – and what was your role in that project?
• What is your relationship with Linux?

2nd round: an (online) interview

Company information

Medgical

Medgical.ai

Introductory paragraph

Clinical trial recruitment is a major bottleneck, with high costs and delays stemming from the manual screening of patients against complex eligibility criteria. The goal of this work is the design, development, and validation of a novel hierarchical matching system. The system employs a cascade of techniques – from simple rule-based logic to advanced Large Language Models (LLMs) – to automatically and efficiently identify eligible patients from electronic health records (EHRs). The primary objectives are to significantly reduce recruitment costs, increase matching precision, and enable real-time integration with clinical systems, thereby accelerating clinical research.

Approximately 80% of clinical trials fail to meet enrollment deadlines, and 30% of trial sites enroll zero patients. The core problem is the reliance on manual, time-consuming, and error-prone processes for screening patient records against lengthy and complex trial protocols. This creates a critical need for automated, accurate, and scalable solutions that can parse both structured and unstructured clinical data to identify suitable candidates efficiently.

Image/video

Video

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Primarily remote. The core work—algorithm development, data analysis, and system design—can be conducted effectively from any location with a secure internet connection.

Format: The default mode is working from a distance. However, we strongly encourage and will facilitate regular in-person collaboration sessions in Tartu. We will coordinate with the student's course schedule to plan these sessions, which could include:
• Initial onboarding and requirement workshops.
• Bi-weekly or monthly progress review and planning meetings.
• Key milestone presentations and testing sessions.

Language: English. All project documentation, code, meetings, and written communication will be in English to fully include international students and align with the technical nature of the work.

Expected general assignments

The student will engage in a full-cycle data science and software engineering project, including:

• Conducting requirements analysis by reviewing clinical trial protocols (de-identified) and interviewing clinical research coordinators to define precise matching logic.

• Independent research and analysis of existing patient matching algorithms, NLP techniques, and LLM application frameworks in healthcare.

• Data pipeline development, including writing scripts for extracting, anonymizing, and preprocessing structured (ICD codes, lab values) and unstructured (clinical notes) data from provided sample datasets.

• Implementing and testing the three-tiered matching algorithm (rule-based, semantic, LLM) in Python, including writing comprehensive unit and integration tests.

• Participating in weekly sync meetings with the academic and industry supervisors to present progress, discuss challenges, and plan next steps.
  

Expected topics for master’s thesis

Research question 1: Can a hierarchical matching architecture, escalating from rule-based to LLM-driven methods, optimize the trade-off between computational efficiency and screening accuracy?

Research question 2: What are the definitive performance metrics (e.g., sensitivity, specificity, speed) for each tier of such a system when applied to real-world clinical trial criteria?

Hypothesis: A hierarchical system will correctly identify eligible patients with >95% sensitivity and >90% specificity while reducing screening time by over 50% compared to manual methods, with over 70% of cases resolved without invoking LLMs.

(Technical) expectations for applying student

Programming skills in Python

Supervisor designated by the partner

Primary supervisor: Dr Armando Vieira

Clinical co-supervisor in Trial Methodology: Dr Diogo Silva

Why should you join us?

You will work with real, de-identified clinical datasets from hospitals and industry partners in a secure research environment with GPU access for LLM development. The project is embedded in an international collaboration, including a pharmaceutical company and The Royal Marsden in London, offering rare exposure to real-world clinical research and high-impact AI applications.

Application process

In addition to your CV, answer the following question:

• Why are you interested in working with Medgical, and how do you see your background contributing to our mission?

• Tell us about a technical project you’ve worked on that you’re particularly proud of. What was your role, and what was the impact?

• Have you worked with NLP, LLMs, or healthcare-related datasets before? If so, describe your experience briefly. If not, how would you get up to speed?

• Medgical values innovation that has measurable real-world impact. Describe a situation where you had to balance technical complexity with usability or real-world constraints.

2nd round: a video interview with the company representatives

Omniva

Website

Introductory paragraph

Omniva is an international logistics company operating worldwide. Omniva is also the national postal service provider of Estonia and operates commercial networks in Latvia and Lithuania.

Omniva delivers goods purchased online – even from distant countries – quickly to customers’ homes, the nearest post office, or to our customers’ preferred destination: 24/7 parcel lockers across the Baltics. We are the largest service provider in the Baltic region.

We understand local markets and stay close to our partners to find the best solutions for each specific region.

Omniva’s international business is a driver of cross-border e-commerce in the Baltics, Eastern Europe, the Caucasus and Central Asia. Our current logistics centres operate in the United States, the United Kingdom, Germany, the Netherlands, Lithuania, Estonia, Finland and Oman. Operations in Kyrgyzstan will begin in September.

We are working purposefully to become the fastest-growing company in the region. We are bold enough to enter new markets and introduce modern solutions there. We are prepared to adapt to all kinds of circumstances and last-minute changes. Our solutions create opportunities to generate more revenue with lower investment. We offer extensive logistics networks and collaborate with the most reliable partners.

Omniva currently employs approximately 2,500 people, all dedicated to ensuring customer satisfaction and bringing the world to customers’ doorsteps.

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tallinn

Format: Hybrid

Language: English

Expected general assignments

• Analyse historical pricing, demand and volume data to understand price sensitivity and demand patterns
• Design and develop simulation models to evaluate the impact of different pricing strategies on revenue, volumes and capacity utilisation
• Develop pricing optimisation models that account for network throughput, service levels and competitive constraints
• Translate simulation results into practical and actionable pricing recommendations in collaboration with business stakeholders

Expected topics for master’s thesis

• Simulation-based evaluation of dynamic pricing strategies in parcel logistics – modelling how different pricing mechanisms (time-based, demand-based, capacity-based) affect revenue and network utilisation
• Multi-objective optimisation in parcel pricing under operational constraints – balancing competing objectives (revenue, volume targets, capacity smoothing, customer fairness) while considering real logistics constraints
• Price elasticity modelling and demand response in B2B parcel services – assessing pricing responses across customer segments and applying these results to pricing rule design

(Technical) expectations for applying student

• Strong foundation in statistics, probability theory and mathematical optimisation
• Strong programming skills in Python (experience with SciPy, OR-Tools, SimPy or similar tools is an advantage)
• Good SQL skills and experience working with large datasets
• Understanding of simulation methods (Monte Carlo, agent-based modelling, discrete event simulation)
• Interest in logistics, pricing or revenue management
• Ability to present and explain results to non-technical stakeholders

Supervisor designated by the partner

Why should you join us?

By joining Omniva, you will have a unique opportunity to contribute at a time when the company is undergoing significant transformation – evolving from a traditional postal company into an international technology-driven enterprise.

As part of our organisation, you will gain insight into how our data team is structured and the challenges they face on a daily basis.

Application process

In addition to your CV and motivation letter, answer the following questions:

• Why would you like to join Omniva?
• What do you hope to gain from this experience?
• What has been the most exciting project you have worked on? Why that one?

2nd round: a (video) interview

Pipedrive

Website

Introductory paragraph

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tartu

Format: When you begin at Pipedrive, we expect your presence in the office. We hold the belief that face-to-face communication and physical presence are more effective during your initial months. Later on, it becomes possible to work remotely once you have gained confidence.

Language: English

Expected general assignments

There are three positions available:

1. AI Engineering position
2. Machine Learning & Data Science position
3. Software Engineering position

No matter the position, you will be a part of a team, participate in planning, standups, and retro meetings.

We don't expect but encourage you to come up with your own solutions, ideas, research findings.

Expected topics for master’s thesis

Pipedrive has multiple initiatives to use Machine Learning for customer-facing features and to improve internal processes. Topics for this year will revolve around:

• Analyzing Customer Success Patterns - what are the trends and clusters of companies who convert from trial to pay based on company size, behaviour, industry, actions?
• Personalised Customer Onboarding - building an engine that provides tailored onboarding recommendations for customers on what to start with
• Smart(er) Knowledge Base Agent - applying advanced personalisation techniques, improving RAG-pipeline for Knowledge Base agent to be proactive and personalised assistant

(Technical) expectations for applying student

AI Engineering position: We want you to have good Python skills, and to be curious about LLMs and agentic systems.

Machine Learning & Data Science position: You've done courses on Python, SQL, and ML libraries (Pandas/NumPy, scikit-learn), and are willing to learn more in a real team with real tasks.

Software Engineering position: You are comfortable with one programming language, REST APIs, SQL and Git are a big plus.

No matter the position, a good candidate has:

• good communication skills

• the ability to write code in any programming language

• got to be fun to work with

Supervisor designated by the partner

Mykhailo Dorokhov, Senior Engineering L&D Lead

As your supervisor, Mykhailo will take care that you are growing as a specialist and that your goals at university are aligned with what you do at Pipedrive.

You will also get a buddy in the team you'll be working with, who will support you through your journey as a software engineering intern.

Why should you join us?

Application process

We have a simple three-step interview process:

Step 1. Every applicant must take a brief cognitive aptitude test. If you like logic puzzles - you'll love this one.

Step 2. Make a short 2-minute video introducing yourself, describing your previous studies/experience and telling us what made you apply to Pipedrive

Step 3. Interview with our engineers and TA partner. We don't bite, and it's your chance to see Tartu from the 16th floor.

Company information

Reach-U

Website

Introductory paragraph

Reach-U works with customers around the world on unique data-driven projects. For example, analyzing mobile networks and fusing data from mobile, internet, and TV traffic. With our purpose-built technology, our customers can perform business analysis much more interactively than with other tools.

Areas of activity: data analysis platforms for telecommunications and media companies, optimization and profiling algorithms, data exploration user interfaces, and geospatial data.

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tartu

Format: Hybrid work is common (individual remote work + office teamwork).

Language: Our working language is a mix of English and Estonian.

Expected general assignments

• Building tools to assess the quality of customer datasets while maintaining privacy.
• Testing beta and new features of Google Maps and Deck.gl, and assessing the applicability of WebGL in advanced visualization methods.
• Experimenting with new visualization methods for datasets in web browsers
• Designing, coding, and evaluating the performance of algorithms to extract insights from real-life datasets.
• Conducting desk studies of publicly available prior art.

Example of the potential task below
Method(s): Multidimensional clustering and interactive visualization
Example dataset: Usage data of the consumer application that contains A/B test results
Task: define clusters based on how users reacted to A/B tests.
Data structure:
a) user activity log
b) user attributes from the customer database (e.g. home city, preferred language etc).

We will start with a reasonably-sized dataset to build a demonstrator. The real dataset includes millions of households and hundreds of millions of view facts.

Optionally, make attractive interactive visualization of the clustering result.

Expected topics for master’s thesis

• Identifying patterns in geographic and behavioral datasets.
• Detecting hotspots, changes, or anomalies in spatio-temporal datasets.
• Developing interactive visualization methods for large datasets in web browsers.

(Technical) expectations for applying student

• Experience in Airflow, JavaScript, Java, Python, or R.
• Preferably experience with Jupyter, Tableau, QlikView, or similar tools.
• Familiarity with at least the basics of statistical analysis and ML methods.

We don’t expect you to know it all; having a few of these skills is more than enough.

Supervisor designated by the partner

Why should you join us?

Reach-U is a 35-year-old University of Tartu spin-off. While we initially focused on geographic information systems and later on mobile operators, today we serve the largest media company in North America.

We currently have several 'mission impossible' projects under delivery, and you can contribute by testing various new approaches. If they succeed, the impact will be huge.

Our team is a mix of back-end and front-end developers, data engineers, and UX/UI designers, all working together to create tools that serve people around the world.

We believe in teamwork. You will be a full member of our team, but if you have a friend or partner you enjoy working with, we are happy to define a task where you can work together as a mini team.

Application process

In addition to your CV, answer the following question:

1. Please describe your experience in a free-form format (feel free to include links to your portfolio or GitHub if you have them).
2. Briefly describe your motivation – ideally in a way that’s convincing and engaging. In which specific areas do you want to develop yourself?
3. Reflecting on your hobby projects or university courses, can you highlight any “Wow!”, “Eureka!”, or “I really enjoyed that!” moments? If so, please describe them.

Based on your answers, we will put together a shortlist of candidates we’d like to meet. The goal is to see if we’re a good match. Perhaps you also have a question or three for us?

Company information

Soldera

Website

Introductory paragraph

Soldera is an AI-native climate-tech startup building autonomous, AI-powered infrastructure that enables the buying, selling, and management of renewable energy certificates across more than 30 European national registries through a single unified platform.

Field of activity: artificial intelligence, energy software, data platforms, automated trading systems.

Image/video

Video

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tartu

Format: On-site

Language: English

Expected general assignments

• Designing and developing AI agents capable of autonomously operating across different national renewable energy certificate registries
• Building data pipelines to ingest and normalize unstructured data (PDFs, scanned documents, heterogeneous API formats) using LLMs and OCR
• Contributing to the architecture of a unified API that enables energy asset managers to operate across multiple countries
• Prototyping, developing, and iterating AI-powered product features in close collaboration with the founders
• Applying LLMs, vector databases, and agentic workflows in a production environment

Expected topics for master’s thesis

• Applying agentic AI systems in complex regulatory environments
• LLM-based normalization of unstructured data in multi-national information systems
• Architecture and reliability of AI-driven trading systems
• Use of RAG and agentic workflows in energy software products

(Technical) expectations for applying student

• Very strong programming skills (TypeScript, Python)
• Prior experience with developing software used in production environments
• Knowledge of LLMs, RAG solutions, and agentic workflows (e.g. LangChain / LangGraph)
• Experience with databases and backend systems (e.g. PostgreSQL)
• Willingness to use AI-assisted development tools (e.g. Copilot, Cursor)
• Independent working style and ability to connect technical solutions with business objectives

Supervisor designated by the partner

Why should you join us?

Soldera offers the opportunity to contribute to a highly ambitious AI-powered product with real impact on Europe’s energy security and green transition. The student will gain hands-on experience in a fast-growing deep-tech company, working on state-of-the-art agentic AI systems and data platforms together with experienced engineers and founders.

Application process

In addition to you CV, please also provide:

• LinkedIn profile
• GitHub profile
• Olympiad results
• Links to personal development projects

Company information

Swedbank

Website

Introductory paragraph

Swedbank is one of the largest banks in our home markets in the Baltics and Sweden, with more than 7.3 million private and 600 000 business customers. We offer a wide selection of financial services and work every day to support people, businesses, and society in growing by promoting a healthy, sustainable economy.

Field of activity: financial services

Image/video

Video

Arrangement of collaboration:

• Location
• Format
• Language requirements

Location: Tallinn & Tartu

Format: Hybrid working, with regular physical presence encouraged

Language: English

Expected general assignments

Our data science team is based within the Anti-Financial Crime (AFC) function of the bank, developing high-quality, reliable data science products. In our team, you will:

• Participate directly in one of our delivery teams delivering products for AFC or the wider bank
• Experience the Scaled Agile Framework for Enterprise (SAFe) way of working
• Collaborate closely with our stakeholders to learn about the domain and ensure the products we build together are relevant to the bank’s needs
• Apply state-of-the-art AI methods to extract knowledge and identify solutions from vast quantities of data
• Contribute to developing, deploying and operationalising data science products

Expected topics for master’s thesis

• Investigate and design an automated evaluation framework for generative AI applications, focusing on user feedback mechanisms and continuous learning to close the feedback loop.
• Assess and apply community-detection and node-similarity algorithms to complex networks using graph database technology, evaluating their effectiveness at uncovering hidden relationships.

(Technical) expectations for applying student

• Team player who lives the Swedbank values of Open, Simple and Caring
• Strong and humble communicator
• Knowledgeable about the application of statistical methods, AI and machine learning to complex data sets
• Familiarity with Python and common development practices (e.g. PEP 8, version control, testing)
• Previous hands-on experience with generative AI and graph database technologies is meriting.

Supervisor designated by the partner

Why should you join us?

Swedbank is over 200 years old and one of the largest banks in each of our home markets, including Estonia. Our customers trust us to handle their finances, but we also have an obligation to society to prevent bad actors from abusing the financial system. Come join the largest data science team at Swedbank and help us build products that protect society and bring the promise of AI to the whole bank.

Application process

In addition to you CV, please also answer the following questions, either in text or video:

• Why do you want to work at Swedbank?
• Who or what inspired you to work in data science?
• In two or three sentences, describe your favourite data science project that you have been involved in, focusing on why the project was important and the tangible outcomes of your work towards that goal.
• What do you want to achieve and learn working with Swedbank during your participation in the Digilink programme?

2nd round: a series of virtual interviews to learn more about you and to test your data science skills

We have a dedicated teamwho promotes and leads the innovations, we name this team E-Lab. The team consists of highly skilled specialists in the field of software engineering, data science and cloud engineering. We test diferent solutions in cooperation with other units and external partners in the framework of short-term prototyping projects. Data Scientists participate in E-Lab as core team members, providing their expertise andknowledge on machine learning, data analysis and optimization.

The Machine Learning engineer, would work closely with data scientists in orderto deploy theirmodel. They would also work closely with software engineers who deal with back-end and infrastructure.

ML engineer would work starting from the feature engineering phase, creating pipelines for data cleaning, transforming and then training and predicting for the ML model in production.

• Share your code
• Present your work.
• Integrate MachineLearning models.
• Use MLOps tools

• Good at machine learning
• MLOps
• Strong IT skills
• Presentation skills
• Interest for energy industry

• AI is more and more usedin the electricity sector.
• Strong team with several data scientists and developers, so we learn a lot from each other.

Questions:

• Why would you be a good fit for this position?

For the second round of admission, there will be a work assignment, as well as an interview with the supervisor.

We have a dedicated teamwho promotes and leads the innovations, we name this team E-Lab. The team consists of highly skilled specialists in the field of software engineering, data science and cloud engineering. We test different solutions in cooperation with other units and external partners in the framework of short-term prototyping projects. Data Scientists participate in E-Lab as core team members, providing their expertise and knowledge on machine learning, data analysis and optimization.

Data scientists work on diverse themes: renewable electricity production forecast, consumption forecast, electricity price forecast, data analysis for electricity production, for electric vehicles, energy storage…

• Create features to make better predictions.
• Deliver data analysis.
• Share your code
• Present your work.
• Finally, integrate your model with developers.

• Good at mathematics, especially probability
• Deep learning
• Presentation skills
• Interest for energy industry

• Interesting topics. You can apply your skills to an industry that matters: energy. Due to more & more renewable energy, electricity production is less known and requires AI to make the system stable.
• Strong team with several data scientists, so we learn a lot from each other.

Questions:

• What kind of model wouldyou try fortime-series forecasting. For example, predicting electricity pricesof the next12 hours.
• Why would you be a goodfit for this position?

For the second round of admission, there will be a work assignment, as well as an interview with the supervisor.

Company description

Tilde is one of the leading European language technology companies. With a team of over 150 professionals across Latvia, Lithuania, and Estonia, we’re proud to develop impactful AI-driven solutions that enable multilingual communication in Europe and beyond.

Website

https://tilde.ai/

Location

Remote or Riga office

Language

Estonian and English

Expected assignments

• Dataset curation and synthetic data generation
• Reproducing and improving previously published research
• Large language model adaptation
• Clear and transparent documentation of results

Topics for master’s thesis

LLM adaptation for Estonian in-context question-answering for RAG-use case

Expectations for applicant

• Hands-on experience with Python and Linux command line proficiency
• Familiarity with commercial LLM APIs
• Understanding of neural network architectures, training techniques, and fine-tuning methods

Questions for the Applicant

1. What is your interest in this project?
2. How do you see your work in this project contribute to your broader career goals, and how do you plan to leverage this experience to grow and develop professionally?
3. Share details about a machine learning project you are especially proud of. What was the project’s focus? What obstacles did you encounter, and how did you address them? If available, could you provide a link to its GitHub repository or any other relevant open-source repository?
4. Describe what skills and tools you might need in this project. Indicate which skills you possess and which you would need to learn within this project.

Supervisor

Toms Bergmanis or Mārcis Pinnis

Why join us?

As an AI Engineering intern specializing in LLMs at Tilde, you’ll play a key role in developing state-of-the-art AI technologies. You’ll collaborate with a multidisciplinary team of researchers and engineers to design, implement, and optimize LLMs that drive Tilde’s language solutions.

Company description

Telia is the largest IT and telco company in Estonia, being market leader in all its business areas. Our mission at Telia is to shape the future of one of the world’s most connected regions, pioneering ways for businesses and societies to become smarter and more sustainable.

We value collaboration, continuous learning, and competent teams. We have a friendly and supportive work environment where each of us does our job with great devotion. The rapid and continuous development of our area also gives us endless opportunities for constant growth and learning.

Learn more: https://www.telia.ee/ettevottest/karjaar-telias and https://www.youtube.com/watch?v=CBKTdIr7H74

Telia as an employer:

Image

Introduction

Join our team as a Network Solutions Architect Trainee

Are you passionate about network solutions and eager to develop your skills in a dynamic environment? We are offering a unique opportunity for training and growth as a Network Solutions Architect. Join us and be part of a team that values innovation, collaboration, and professional development.

Location

Office in Tallinn (Mustamäe tee 3) or Tartu (Turu 26B)
Hybrid work is a common way of working.

Language

Basic knowledge of Estonian is mandatory.

Expected assignments

• Map and model the architecture of existing network inventory systems (as is).
• Develop and model target architectural solution for network inventory (to be architectural vision): Solution design, Integration planning, Scalability and Performance, Security, and Compliance (GDPR, ADR’s), Estonia National Fallback Service requirements.
• Investigate and identify suitable products available on the market for solution implementation.
• Work closely with OSS/BSS teams to integrate inventory data with operational systems.
• Support regulatory compliance and audits related to network assets.
• Understanding of network asset lifecycle management

Topics for master’s thesis

Futureproofing Network Inventory

Expectations for applicant

Technical Skills (while expertise in all listed technologies is beneficial, a strong understanding of 70-80% of them is sufficient.)

• Understanding of IP/MPLS networks, fiber optics, transmission (DWDM, SDH), mobile networks (4G/5G)
• Experience with operating systems (Windows, Linux)
• Familiarity with firewalls and security protocols
• Knowledge of APIs and data integration (REST, SOAP, XML, JSON)
• Some experience in Java & Microservices (Kubernetes)
• SQL, PostgreSQL, MySQL for data querying and reporting
• Experience with GIS tools (ArcGIS, QGIS) for fiber mapping - is plus not not expected.

Analytical and Problem-Solving Skills

• Good analytical and conceptual skills for designing and implementing IT solutions.
• Ability to solve technical problems and provide innovative solutions

Knowledge of Technology and Trends

• Continuous interest and awareness of current and future technologies and trends.
• Ability to recommend and implement new technologies based on company needs.

It is beneficial to have at least one certification:

• Cisco Certified Network Associate (CCNA)
• CompTIA Network+
• Microsoft Certified: Azure Network Engineer Associate
• AWS Certified Advanced Networking.

Supervisor

Ardo Abel, Domain Architect

Why join us?

Join us to see and experience life in one of the largest IT teams in Estonia, where we develop and manages approx 800 systems (from legacy to innovation). As part of your internship, you will have the chance to learn about our IT and network systems, architecture, and work with different experts in the field.

Application process

We start with CV/LinkedIn profile and motivation letter. Questions to address in text or through video

1. What inspired you to apply for the position? Why would you be a good fit for this position?
2. Please describe your experience with network technologies and services. If you have participated in implementation projects, then describe your role.
3. How do you stay updated with the latest trends and technologies in network architecture?
4. What do you want to achieve and learn working with us at Telia during your industrial masters?

If we take your application further, we will have a follow-up interview to learn more about you and your motivation.

Company description

MyDello is revolutionizing freight forwarding by bringing global freight online. Check out our website at MyDello.com

Introduction

MyDello, launched in 2021, is fast-growing digital freight forwarder that offers the best shipping options across all freight modes, providing instant quotes and guaranteed prices. We believe that global logistics will become fully automated, and MyDello is on a mission to connect all parties in the supply chain to significantly reduce overall logistics costs.

Introduction video

Video

Location

Our office is located in Tallinn, but we support and encourage remote work. However, we believe face-to-face communication in the office is crucial for success during the initial settling-in period.

Language

English or Estonian

Expected assignments

The specific tasks depend on agreements made, but generally, you will:

• Work in the development team and participate in:
  • daily meetings,
  • planning new functionality,
  • researching requirements.
• Work on and experiment with your own development ideas.
• Document your findings and work.
• Test your code.
• Research and identify the science proved methods to calculate CO2 emissions in one of the most polluting industries and model the ways to reduce CO2 emissions.
• Design and build scalable data model to implement the identified approach.
• Collaborate with the design and front-end development team to create meaningful CO2 emissions representations for customers in the MyDello customer portal.

Topics for Your Master’s Thesis

Expectations for applicant

• Knowledge and experience with statistical analysis and the ability to gather meaningful data to make quality decisions.
• Experience in research and finding quality studies to solve a problem.
• Understanding the basics of back-end development (C# and .NET framework) is a plus (we provide support from senior engineers).

Supervisors

Rasmus Karja, Product owner at MyDello
Siimu Kaas, CTO and Co-founder at MyDello

Why should you join us?

Join us and lead the future of digital logistics. We serve more than 1,500 customers across Europe. Your science-backed input will have a meaningful role in digitalization of the $10T logistics industry.

We value teamwork, creativity, and learning from mistakes. You will join a passionate team where your ideas matter. You will be a full member of our amazing global team from day 1.

Application process

Our process is simple:

• Step 1: Describe your experience in free form and give a short overview of how you can commit to our journey. Please also choose which topic feels more interesting for you.
• Step 2: Attend an initial interview with our HR manager to determine if both sides see a strong fit for collaboration.
• Step 3: Participate in a final interview with MyDello CTO and product owner. If we find the spark, you’ll receive an offer to join our team.

Company description

MyDello is revolutionizing freight forwarding by bringing global freight online. Check out our website at MyDello.com

Introduction

MyDello, launched in 2021, is fast-growing digital freight forwarder that offers the best shipping options across all freight modes, providing instant quotes and guaranteed prices. We believe that global logistics will become fully automated, and MyDello is on a mission to connect all parties in the supply chain to significantly reduce overall logistics costs.

Introduction video

Video

Location

Our office is located in Tallinn, but we support and encourage remote work. However, we believe face-to-face communication in the office is crucial for success during the initial settling-in period.

Language

English or Estonian

Expected assignments

The specific tasks depend on agreements made, but generally, you will:

• Work in the development team and participate in:
  • daily meetings,
  • planning new functionality,
  • researching requirements.
• Work on and experiment with your own development ideas.
• Document your findings and work.
• Test your code.
• Design and build functionality to retrieve shipment tracking data from various carriers via APIs or other channels, using LLM tools to extract meaningful data for the MyDello shipment tracking system.
• Collaborate with the MyDello core development team to implement the functionality into the core system.

Topics for Your Master’s Thesis

Expectations for applicant

• Experience or strong interest with LLM methods and tools.
• Knowledge of gathering and storing data from different sources.
• Understanding the basics of C# and the .NET framework is a plus (we provide support from senior engineers).

Supervisors

Rasmus Karja, Product owner at MyDello
Siimu Kaas, CTO and Co-founder at MyDello

Why should you join us?

Join us and lead the future of digital logistics. We serve more than 1,500 customers across Europe. Your science-backed input will have a meaningful role in digitalization of the $10T logistics industry.

We value teamwork, creativity, and learning from mistakes. You will join a passionate team where your ideas matter. You will be a full member of our amazing global team from day 1.

Application process

Our process is simple:

• Step 1: Describe your experience in free form and give a short overview of how you can commit to our journey. Please also choose which topic feels more interesting for you.
• Step 2: Attend an initial interview with our HR manager to determine if both sides see a strong fit for collaboration.
• Step 3: Participate in a final interview with MyDello CTO and product owner. If we find the spark, you’ll receive an offer to join our team.

About Montonio

Montonio is an e-commerce checkout platform, loved by over 8000 merchants in the Baltics, Finland, and Poland. To date, we have raised nearly 19M € from Index Ventures and the people who built Wise, Bolt, and Pipedrive. We aim to be the go-to partner online merchants can rely on to grow their business, one superb checkout experience at a time. 🚀

https://montonio.com/

Introductory paragraph

As a software engineer at Montonio, you’ll be able to dive into account-to-account payments, the magic of e-commerce checkout experiences, parcel shipping, anti-money laundering, or other exciting projects. With a full-stack role, you'll tackle both front-end and back-end tasks, contributing to the development and refinement of user experiences.

Image/video

Read more about the position and our culture here

Arrangement of collaboration:

• Location
• Format (physical presence vs working from a distance)
• Language requirements (Estonian/English/either-or/both)

You'll be working from our Tallinn office, where our main engineering team is based. During the first 4 months, we expect you to work on-site most days to get properly integrated into the team. After this initial period, it is possible to transition to a hybrid arrangement with 1-2 remote days per week, although we value and encourage in-person collaboration.

Our working language is English, used for all communication and documentation.

Expected general assignments

• Work with your team to generate impactful ideas and proposals for features and product enhancements.
• Partner with your fellow engineers to turn these ideas into reality, build great user experiences, and iterate.
• Explore and learn new technologies that help us build better products for our merchants.

Expected topics for master’s thesis

As a startup operating in a fast-changing environment, it’s difficult to foresee the projects you will be working on. However, here’s a selection of potential topics:

• Analyzing and optimizing the customer onboarding flow
• Implementing user-friendly and efficient anti-money laundering measures in e-commerce
• Analyzing and implementing anti-fraud solutions for card payments
• Developing a product recommendation engine

(Technical) expectations for applying student

Our expectations for you:

• You're comfortable with at least one modern JavaScript front-end framework (Angular, React, Vue, etc.) and are proficient in HTML and CSS.
• You have experience with back-end development in either TypeScript, JavaScript, or Java. You have experience with using relational databases.
• You care about how and why users interact with the product.
• You’re a skilled communicator – you’re proficient in written and spoken English and are able to work cross-functionally with designers, product managers, and other engineers.
• You work great individually, but even better with a team that you can depend on and learn from.
• You are curious, adaptable, and open to challenges.

Bonus points:

• Experience with modern cloud-native software development practices and architecture patterns.
• Track record of taking ideas to execution – from development through deployment to post-launch support.

We believe passionate engineers can thrive in any environment, no matter the language or framework. Hence, there’s no need to be already familiar with all of our tech. Here are some of the technologies you can expect to work with at Montonio:

• Node.js and TypeScript on the back-end (NestJS framework).
• Angular on the front-end.
• Our database of choice is PostgreSQL.
• We run our code in Docker containers and make heavy use of the AWS infrastructure.

Supervisor designated by the partner

Kristofer Turmen, CTO & Co-Founder

Why should you join us?

We’re all about creating products with heart and soul. As a Montonio software engineer, your focus will be on coding and collaborating to craft compelling experiences. You'll play an integral role in the product lifecycle, from design to deployment, incorporating user feedback for iterative improvements. Your dedication to our merchants' success and understanding of our product's broader context will be key to your success. True end-to-end ownership is the name of the game in Montonio.

As a young team, we don’t have strict processes or archaic structures. As we’re still building our engineering culture, your input will help us create an environment that drives innovation while giving people the space to try out new things. There are plenty of opportunities to experiment and contribute to our shared goal of delivering exceptional products while supporting the needs of our customers and team.

You would be joining Montonio at a very opportune time. We’re still a team of 60 people and we are growing. It is a very exciting time that enables the possibility to be part of shaping the organization, our products, and Montonio's brand. You’ll be part of a kickass team of friendly, forward-thinking people who always support each other. ❤️

Application process

• What caught your attention with Montonio?
• Tell us of a project that taught you the most.

Here is our application process. Not all candidates will progress to all stages.

1. Initial screening interview (video call)
2. We send you our home task
3. Home task review and whiteboarding (on-site)
4. Team interview (on-site)
5. Offer (video call)

Company description

Veriff

Website: https://www.veriff.com

Introduction

We are the preferred identity verification platform partner for the world’s most innovative growth-driven organizations helping conveniently verify and safeguard users anywhere in the world. We support the broadest number of identity documents from nearly every country and territory in the world - and this is continually increasing!

We build ML models that safeguard honest people online. As an intern, you’ll join our ML teams in creating the next generation of AI-driven identity verification solutions.

Take a look at this interview with Krister, one of our previous interns (and current colleague): https://www.veriff.com/our-people/tech-talent-krister-kasemaa

Arrangement of collaboration:

Location: Tallinn (hybrid is possible)
Format: In-person and remote mix (with the expectation of some on-site teamwork)
Language: English is our main working language.

Expected general assignments

• Develop and validate PoC solutions for ML-driven features
• Work collaboratively by joining planning sessions, daily standups, and retrospectives
• Contribute to ML Model development and testing in cloud-based development environment
• Coordinate with cross-functional teams to integrate updates and drive seamless collaboration
• Maintain clear, transparent documentation of results

Expected topics for master’s thesis

• Biometric Verification: Explore novel machine learning architectures to enhance speed, accuracy, and reliability in verifying user identities.
• Biometrics: Investigate the latest deep learning approaches for face recognition, focusing on real-world performance and scalability.
• Advanced Spoofing Attack Prevention: Develop and assess advanced spoof detection mechanisms (e.g., liveness checks) to reduce vulnerabilities in biometric systems.
• Fairness & Bias Assessment and Mitigation: Evaluate how demographic differences impact model outcomes, and propose strategies to reduce bias in biometric verification.

(Technical) expectations for applying student

• Understanding of machine learning concepts
• Familiarity with Python and common data science libraries (NumPy, Pandas, TensorFlow or PyTorch)
• Familiarity with SQL
• Team player with a proactive approach to problem-solving

Supervisor designated by the partner

Doğuş Karabulut, Senior Engineering Manager

Why should you join us?

At Veriff, you’ll be at the heart of machine learning innovation. You’ll gain hands-on experience shaping AI that protects millions of people globally. We give you the freedom to learn, grow, and contribute to real-world products that make a difference.

Application process

Round 1: Short video where you introduce yourself and share why you want to join us at Veriff.

What excites you about applying deep learning to identity verification?

Describe a small project you’ve done and your biggest takeaway.

How do you see this internship contributing to your long-term career goals?

Round 2: Virtual or in-person interview with our Data Science team. We’ll discuss your ideas, a potential thesis topic, and walk you through the day-to-day internship tasks.

Videos:

• We are Bolt
• Come build with us

Photos

Website

https://medium.com/bolt-labs

Business domain

Mobility

Location

Tartu

Languages required

English

Expected assignments

• Implementation of proof of concept (PoC)
• Reading, presenting, and reproducing previously published research related to the problem.
• Clear and transparent documentation of results

Topics for master’s thesis

We have different directions depending on the candidate's background and interest:

1. Machine learning models for long-term effects estimation in AB tests.
2. LLM applications for knowledge-specific tasks in AB test platforms.

Expectations for applicant

• Experience with Python and data related data science libraries (Numpy, Pandas)
• Familiarity with SQL and Jupyter Notebooks
• Knowledge/interest in experimentation and causal inference topics.

Questions for the Applicant

1. What drew you to the specific topic you are applying for?
2. How do you see your work in this program contribute to your broader career goals, and how do you plan to leverage this experience to grow and develop professionally?
3. Describe a machine learning project you've worked on that you are proud of. What was it about? What were the challenges you faced, and how did you overcome them?

Supervisor

Carlos Bentes, Senior Data Scientist

Why Bolt?

We value people's potential over experience. We believe that talented and hard-working people grow quickly, so we give them opportunities that most other companies would not.

The experimentation platform is a cross-functional distributed team of data scientists and software engineers that work on exciting problems with an impact, helping Bolt transform urban transportation and make cities for people, not cars.

Find more about Bolt:

• https://bolt.eu/en/careers/life-at-bolt
• https://bolt.eu/en-ee/cities/solutions

Company description

Pipedrive

Software development (CRM & intelligent revenue platform)

https://www.pipedrive.com/

Introduction

We are confident that Pipedrive needs no introduction. As the first CRM to employ Kanban for visualizing sales, Pipedrive was established in 2010 by five Estonian engineers and entrepreneurs. By 2020, it had become the fifth Estonian company to achieve unicorn status.

Image/video

Video

Arrangement of collaboration:

• Location
• Format (physical presence vs working from a distance)
• Language requirements (Estonian/English/either-or/both)

Location: Tartu

Language: English

Format: When you begin at Pipedrive, we expect your presence in the office. We hold the belief that face-to-face communication and physical presence are more effective during your initial months. Later on, it becomes possible to work remotely once you have gained confidence.

Expected general assignments

• Software development and testing

• Working in a cloud-based development environment

• Working in a team:

  • taking part in planning, standups, retro meetings

  • Cross-team collaboration

• We don't expect but encourage you to come up with your own solutions, ideas, research finding

Expected topics for master’s thesis

Pipedrive has multiple initiatives that aim to utilize Machine Learning to improve customer-facing, as well as internal parts of our system:

• Applying machine learning to different parts of our product;

• Continuous integration and continuous delivery metrics;

• Automations in site reliability engineering;

• Using RAG, vector databases and semantic search to support internal processes.

But of course, there are also many other topics that could be available depending on your interests.

(Technical) expectations for applying student

• Good communication skills

• Ability to write code in any programming language

• Be fun to work with

Supervisor designated by the partner

Mykhailo Dorokhov, Senior Engineering L&D Lead

As your supervisor, Mykhailo will take care that you are growing as an engineer and that your goals at university are aligned with what you do at Pipedrive.

You will also get a buddy in the team you'll be working with, who will support you through your journey as a software engineering intern.

Why should you join us?

Pipedrive is one of the pioneers that joined the program back in 2017 and has been participating ever since. We recognize the potential of what a master’s student can achieve and really appreciate the value of a good thesis. Many of the program’s alumni are currently our employees, continuing their work at Pipedrive.

Application process

We have a simple three-step interview process:

Step 1. Every applicant must take a brief cognitive aptitude test. If you like logic puzzles - you'll love this one.

Step 2. Make a short 2-minute video introducing yourself, describing your previous studies/experience and telling us what made you apply to Pipedrive

Step 3. Interview with the internship manager, our engineers and TA partner. We don't bite, and it's your chance to see Tartu from the 16th floor.

Company description

Starship Technologies is the world-leading robotic delivery company, operating in more than 100 cities across the world. We design, build, and operate fleets of delivery robots and use these to provide delivery services to grocery retailers, restaurants, and delivery apps.

https://www.starship.xyz/

Introduction

Our engineering teams cover a wide range of domains, from the robot’s software, hardware and communications with servers, through optimization and supply-demand matching tools, to apps for customers, merchants and operations teams.

Image/video

An example of how we work - overcoming challenges posed by snow

Arrangement of collaboration:

• Location
• Format (physical presence vs working from a distance)
• Language requirements (Estonian/English/either-or/both)

Starship’s engineering teams are located in Tallinn and Helsinki. The more time you spend physically at the office, the more you will likely get out of this collaboration. That said, many engineers prefer to work remotely for at least part of the week, and you will also have the option of working remotely. The working language in Starship’s engineering teams is English.

Expected general assignments

Estimating the battery level of a delivery robot and its depletion rate under various circumstances. Improving the accuracy of battery level prediction of a robot after a driving/idling/charging task would help us make better planning decisions and increase efficiency of the service. In this project, you would:

• build pipelines to gather historical data;
• analyze situations where the existing model is inaccurate;
• iteratively build improved model(s);
• implement the models in production, adhering to efficiency and robustness constraints.

Expected topics for master’s thesis

Estimating the battery level of a delivery robot and its depletion rate under various circumstances

(Technical) expectations for applying student

• Strong quantitative reasoning skills
• Strong basis in programming - your code needs to be performant and resource-efficient
• Ability to adapt to Starship's infrastructure and development environment (Databricks, Kubernetes, Go, Rust etc.)
• SQL and data analysis skills
• Initiative and drive to have an impact

Supervisor designated by the partner

Martin Pihlak, Team Lead of the Core Services team

Why should you join us?

• Improve your data analysis and coding skills
• Get hands-on experience in both data science and cloud services at scale
• Opportunity to deep dive into delivery robot energy management – from battery firmware to fleet action planning.
• Learn from incredibly talented colleagues
• Gain experience in real-world problem-solving, teamwork, and best practices in software engineering
• Help us fill the world with delivery robots

Application process

For the first round of admission, please answer the following questions:

• Why would you like to work on this project?
• Describe specific experiences / projects that set you up for success in this role.

In the second round of admission, you can expect:

• a screening interview
• a technical interview with the supervisor
• a case study, with ~1h to solve a problem, and a discussion afterwards.

Company description

Starship Technologies is the world-leading robotic delivery company, operating in more than 100 cities across the world. We design, build, and operate fleets of delivery robots and use these to provide delivery services to grocery retailers, restaurants, and delivery apps.

https://www.starship.xyz/

Introduction

Our engineering teams cover a wide range of domains, from the robot’s software, hardware and communications with servers, through optimization and supply-demand matching tools, to apps for customers, merchants and operations teams.

Image/video

An example of how we work - overcoming challenges posed by snow

Arrangement of collaboration:

• Location
• Format (physical presence vs working from a distance)
• Language requirements (Estonian/English/either-or/both)

Starship’s engineering teams are located in Tallinn and Helsinki. The more time you spend physically at the office, the more you will likely get out of this collaboration. That said, many engineers prefer to work remotely for at least part of the week, and you will also have the option of working remotely. The working language in Starship’s engineering teams is English.

Expected general assignments

Evaluating a heuristic optimizer for robot logistics planning instead of a genetic optimizer. While our current genetic optimizer works well in practice, it has a drawback of being complex to reason about and modify. A heuristic (rule based) optimizer might be a better choice in some cases.

We would expect you to:

• develop a heuristic optimizer;
• choose metrics to evaluate its goodness;
• run simulations for both optimizer versions;
• evaluate simulation results;
• possibly deploy the new optimizer to a [limited] production and evaluate its real life performance.

Expected topic for master’s thesis

Evaluating a heuristic optimizer for robot logistics planning

(Technical) expectations for applying student

• Strong quantitative reasoning skills
• Strong basis in programming - your code needs to be performant and resource-efficient
• Curiosity about optimization algorithms
• SQL and data analysis skills
• Initiative and drive to have a business impact

Supervisor designated by the partner

Tanel Hiir, Principal Software Engineer in the Fleet Orchestration team

Why should you join us?

• Improve your data analysis and coding skills
• Get hands-on experience with Rust and TypeScript
• Learn from incredibly talented colleagues
• Gain experience in real-world problem-solving, teamwork, and best practices in software engineering
• Help us fill the world with delivery robots

Application process

For the first round of admission, please answer the following questions:

• Why would you like to work on this project?
• Describe specific experiences / projects that set you up for success in this role.

In the second round of admission, you can expect:

• a screening interview
• a technical interview with the supervisor
• a case study, with ~1h to solve a problem, and a discussion afterwards.

Company description

Founded in 2019, Waybiller is an end-to-end bulk transport management solution dedicated to transforming the logistics industry by providing efficient, paperless solutions that simplify workflows, enhance competitiveness, and conserve resources.

Introduction

In our development we are leveraging Django framework on backend and React on frontend, along with CI/CD pipelines in a cloud environment, ensuring a modern, scalable, and high-performance system. We follow the Scrum methodology in our Agile development process, allowing us to continuously iterate and respond to evolving business needs effectively. We are always looking for new and innovative ways to develop and implement technology to make logistics smarter.

Our small but dynamic team comprises top developers, sales representatives, customer success managers, and product designers. Backed by our parent company, Thorgate, we bring expertise, innovation, and mentorship to everything we do. We believe that strong communication and smart collaboration are the keys to our success. Daily standups, business development meetings, and regular feedback sessions keep us aligned and productive.

Our team’s motivation comes from the outstanding results Waybiller produces for its clients and our 3Cs—communication, communication, and communication—keep everything running smoothly.

Website

https://waybiller.com

Location

Tallinn / Remote

Language

English

Arrangement of collaboration

You will begin by familiarizing yourself with Waybiller’s existing systems and understanding how AI can enhance logistics workflows. Initially, you will work on fine-tuning large language models (LLMs) to create an AI Assistant tailored to Waybiller’s use cases.

Once the fundamental AI capabilities are developed, you will move on to designing and implementing workflows and intermediate services that connect the AI Assistant with Waybiller’s backend systems. Throughout the project, you will document architecture, processes, and results in a clear and transparent manner. By the end of this journey, your work will contribute to a fully integrated AI-driven assistant that enhances user experience and operational efficiency in logistics management.

Expected assignments

• Fine-tuning large language models (LLMs) for creating an AI Assistant to be used for available use cases in Waybiller.
• Developing pipelines to process and interpret client inputs, including written and verbal commands.
• Creating workflows and relevant intermediate services to connect the AI assistant with Waybiller’s backend systems.
• Documenting architecture, processes, and results clearly and transparently.

Topics for Your Master’s Thesis

• Fine-tuning LLMs for task-specific natural language understanding in logistics applications.
• Adapting LLMs for effective interaction with logistics platforms.

Expectations for applicant

• A passion for making people’s lives easier through innovative solutions.
• Commitment to continuous learning and staying updated with industry trends.
• Relevant experience with Python and Linux command line.
• Familiarity with fine-tuning open-source LLMs or working with commercial LLM APIs (e.g., ChatGPT / Whisper).
• Experience with RESTful APIs, Docker, and cloud-based AI platforms (e.g., AWS, GCP) is a plus.

Supervisors

Bora Büyükdoluca - Engineering Manager

Bora will focus on your growth and help align your internship goals with your work at Waybiller. You will also have a buddy from the Waybiller Development Team, supporting you with your development process and helping you achieve your goals working on Waybiller.

Why should you join us?

Application process

1. What excites you about developing an AI Assistant for Waybiller?
2. How do you see this project aligning with your broader career goals, and how will it help you grow professionally?
3. Do you have any experience working on an AI/ML project? If yes, please share some details regarding what kind of challenges you faced, and how you solved them? If applicable, provide a link to relevant repositories.
4. What skills and tools do you think are necessary for this project? Highlight which skills you already possess and which ones you aim to develop.

The first round is screening the applications and the answered questions. The answers can be sent either in a written form or as a short video - as long as it is in English!

The second round would be a test assignment based on our tech stack and an interview with the supervisor and HR.

Company description

Mercans - Global Payroll SaaS

https://mercans.com/

Introduction

Mercans is an international IT company offering comprehensive payroll and HR services. We specialize in developing innovative cloud-based Payroll, HR and Recruitment software solutions. Our main product is HR blizz, which is a web-based Payroll & HR system aimed at all level enterprises all over the world. Our goal is to make a payroll and HR system that just works – without the user having to be an expert on taxation and finance.

Our team consists of thriving professionals from diverse backgrounds that all share passion for innovation. We put great effort into fostering a friendly and supportive atmosphere across all our teams, ensuring a workplace where everyone feels valued and empowered.

Introduction video

Mercans' Global Payroll Platform - HR Blizz

Mercans

Location

Tartu, Estonia

We highly encourage interns to work from our office in Tartu. We believe that being physically present in the office provides the best learning experience, facilitates collaboration with our team, and ensures smoother onboarding into our projects. That being said, we are also open for hybrid arrangements based on need.

Language

English

Expected assignments

• Assisting in ML feature engineering, including data preprocessing and transformation for payroll-related datasets
• Supporting data processing tasks, such as data cleaning, visualization, and exploratory analysis
• Contributing to the development and maintenance of ML Ops pipelines, ensuring smooth model deployment, monitoring, and re-training
• Assisting in ML Operations, including deploying models into production and helping refine existing models for improved performance
• Working closely with senior data scientists, data engineers, or ML engineers, occasionally in pairs, to help develop and experiment with new models for payroll automation
• Conducting research on AI-driven enhancements for payroll accuracy and compliance
• Participating in technical discussions, daily stand-up meetings, sprint planning, and backlog updates
• Supporting the development team with coding, debugging, and documentation tasks

The position combines elements of ML engineering, ML Ops engineering, and Data Science. Eventually you will have the opportunity to explore and grow in one or more of these directions.

Topics for Your Master’s Thesis

• Automation of payroll processes using AI-driven workflows
• Enhancing payroll accuracy through NLP-based data extraction
• Designing scalable ML models for international payroll automation and anomaly detection.

Expectations for applicant

• Understanding of ML, AI, and data processing techniques
• Previous experience with Python (preferrably) or Java/Kotlin programming languages
• Understanding of Agile development processes

Supervisors

Oleg Denysenko - Director of AI/ML Engineering
Parag Saxena - Senior Data Scientist
Danica Novakovic - Senior Back-End Developer

Why should you join us?

Mercans combines 20 years of global experience with an innovative, forward-thinking mindset. We are pioneers in payroll technology, constantly pushing boundaries to automate and optimize complex processes.

At Mercans, you’ll find a fun team of professionals from all over the world that thrives on big ideas, collaboration, and innovation. Rest assured, we are committed to providing you with every opportunity to learn, grow and stretch to the highest level of your ability and potential.

Application process

1. Tell us about yourself
   In a short description, share your experience with ML and AI and what excites you about working on payroll-focused ML features. What makes you a great fit for this project and our team?
2. Interview
   Selected candidates will be invited for an interview, either in person or virtually.
3. Test assignment
   Candidates may also be asked to complete a short practical assignment related to payroll-focused ML features to demonstrate their skills.

Company description

Introduction

University of Tartu Autonomous Driving Lab was founded in 2019 together with Estonian ride-hailing unicorn Bolt. The main goals of our lab are:

• Evaluate the self-driving technology for Bolt’s business cases, e.g. robotaxi.
• Create a competence center of self-driving technologies in Estonia and Tartu.
• Prepare future workforce for Estonian self-driving companies, including Bolt.

Our lab has a real self-driving vehicle built on top of the Lexus SUV platform which we regularly test in the city center of Tartu.

Introduction video

Location

Language

Estonian or English

Expected assignments

Testing track:

• Preparing scenarios for testing the self-driving vehicle in simulation
• Mining driving logs for interesting traffic situations
• Converting driving log situations into test scenarios

Machine learning track:

• Training machine learning models for traffic light detection
• Training machine learning models for object detection
• Fusing detections from multiple models / sources

Topics for Your Master’s Thesis

• Scenario mining for testing self-driving vehicles
• Temporal traffic light detection for blinking light detection
• Temporal multi-view/multi-modal object detection

Expectations for applicant

Testing track:

• Comfortable with Python
• Prior exposure to Robot Operating System (ROS) is a plus
• Driving license is a plus

Machine learning track:

• Comfortable with Python
• Comfortable with one of the major machine learning frameworks, e.g. PyTorch, TensorFlow, Keras, etc.
• Statistics/data analysis background is a plus

Supervisors

Tambet Matiisen - project manager

Why should you join us?

Application process

Start by first answering these four questions:

1. What draws you to self-driving cars?
2. What have been the most challenging projects that you have done previously, either in school or for personal use? Please include Github/demo links if possible.
3. What is your experience with robotics and hardware in general? Do you enjoy the messiness of the physical world, or would you prefer to work in a clean simulation?
4. What is your ideal work style? Do you prefer to work in private or in a team? Do you prefer clear guidance o,r given a goal, you prefer to find the solution on your own?

There are no right or wrong answers here; different people are suited for different tasks.

Based on the answers/links we may decide to give you a trial assignment. The assignment will be a small coding exercise in Python.

The final stage is a face-to-face interview with the team.

Company description

Introduction

Introduction images

Location

Language

Estonian or English

Expected assignments

• Collaborate with the team to prepare and analyze unstructured data (financial reports, news).
• Contribute to training and testing machine learning models (LLM, RL).
• Test our experimental software prototype (TRL 7) and validate results.
• Assist with front-end/back-end integration (e.g., connecting data pipelines).
• Help document processes (e.g., how-to guides, test protocols).

Topics for Your Master’s Thesis

• Applying reinforcement learning to optimize stock recommendations.
• Fine-tuning LLM’s for reasoning about stock recommendations.
• Using generative language models for automated financial report summarization

Expectations for applicant

• Know-how in AI, machine learning, and/or data analytics
• Familiarity with Python (preferably with frameworks like TensorFlow or PyTorch)
• Experience in data wrangling, NLP, or RL is a plus
• Willingness and ability to work in a team, communicating openly in English or Estonian

Supervisors

Kristjan Eljand, CEO - Project Manager, AI Strategy Lead, Designer of investment models

Why should you join us?

Application process

First round: Submit a brief motivation letter/video: Why are you interested in AI-driven stock analysis?

1. Include an example or overview of an AI/ML project you’ve done (coursework or hobby).
2. Describe which programming languages and frameworks you’re familiar with (e.g., Python, PyTorch, TensorFlow).
3. Optionally, propose an initial idea for your Master’s thesis.

Second Round: in-person or virtual interview. Possible test task on model testing/code improvement.

Company description

Introduction

Automotive accessories designed, developed, and manufactured by Tarmetec are sold under the Metec brand in most European countries. In addition to our own products, we also offer subcontracting services. Tarmetec's subcontracted production consists of end products used in the automotive industry, machinery manufacturing, healthcare, pharmaceuticals, the food industry, and more.

Our clients are international, publicly listed corporations, for whom we must maintain a high technological level and lead in innovation.

Join us and make your contribution!

Introduction video

Location

Language

Good knowledge of Estonian

Expected assignment

• Participation in meetings
• Preparation of documentation
• Data visualization

Topic for Your Master’s Thesis

The aim of the master's thesis is to compare the theoretical execution times (typically defined in seconds per part) with the actual execution times, analyze possible deviations, and optimize batch sizes to achieve efficiency in production processes and better resource allocation.

To achieve this, the following is required:

• Analyze data accumulated over the past 5 years regarding the start and end times.
• Use statistical methods to compare the differences between theoretical and actual execution times and identify the causes.
• Analyze the impact of batch sizes on the time costs and efficiency of production processes.
• Use analytical or machine learning methods to find optimal batch sizes.

Technical and methodological requirements:

• Data preparation
• Statistical analysis (regression analysis, variance analysis, identifying deviations)
• Machine learning (model development)
• Application of optimization algorithms

The task provides the student with clear objectives and guidelines on how to analyze large datasets, compare theoretical and actual work times, and develop strategies (models) to optimize production processes.

Expectations for applicant

• Knowledge of the CRISP-DM methodology
• Proficiency in SQL at an intermediate level
• Intermediate knowledge of Power BI
• Proficiency in using Jupyter Notebook

Supervisors

Why should you join us?

Questions for the applicant

1. Which tools and methods would you use for data processing and visualization to identify opportunities for improving the efficiency and effectiveness of manufacturing processes?
2. Which methods and tools would you use for collecting and analyzing actual work times to compare them with theoretical work times?
3. What are your strengths that help you handle complexity and challenges?

Company description

Swedbank

Business domain: Financial services

Introduction

Swedbank is one of the largest banks in our home markets in the Baltics and Sweden with more than 7.3 million private and 600 000 business customers. We offer a wide selection of financial services and work every day to support people, businesses, and society to grow by promoting a healthy and sustainable economy.

www.swedbank.ee/careers

Introduction video

Location

Language

English

Expected assignments

Our data science team is based within the Anti-Financial Crime (AFC) function of the bank, developing high quality and dependable data science products. In our team you will:

• Participate directly in one of our delivery teams delivering products for AFC or the wider bank
• Experience the Scaled Agile Framework for Enterprise (SAFe) way of working
• Collaborate closely with our stakeholders to learn about the domain and ensure the products we build together are relevant to the bank’s needs
• Apply state-of-the-art AI methods to extract knowledge and identify solutions from vast quantities of data
• Contribute to developing, deploying and operationalizing data science products

Topics for Your Master’s Thesis

• Develop advanced methodology for creating synthetic data representative of our customers’ behaviors
• Apply federated learning to share model information in collaboration with AI Sweden and other organizations
• Mine and collate information from external sources to better understand our customers and their needs, either through traditional AI methods or Generative AI

Expectations for applicant

• Team player that lives the Swedbank values of Open, Simple and Caring
• Strong and humble communicator
• Knowledgeable about the application of statistical methods, AI and machine learning to complex data sets
• Familiarity with Python and common development practices (e.g. PEP 8, version control; testing)

Supervisors

Why should you join us?

Application process

Questions to address in an application either in text or through video

1. Why do you want to work at Swedbank?
2. Who or what inspired you to work in data science?
3. In two or three sentences describe your favourite data science project that you have been involved in, focussing on why the project was important and the tangible outcomes your work to that goal.
4. What do you want to achieve and learn working with Swedbank during your industrial masters?

If we take your application further, we will have a series of virtual interviews to learn more about you and to test your data science skills.

Company description

Introduction

Apromore is a global provider of AI-driven business process optimization technology. Our AI copilot enables managers to analyze and simulate their processes based on data. Come and join our AI dev team!

https://apromore.com

More information

Location

Language

English

Expected assignments

You will be part of our AI copilot development team, reporting to our tech lead for GenAI. Your tasks will include:

• Write code in Python and/or Java
• Peer-review pull requests
• Write unit tests
• Participate in agile ceremonies
• Participate in solution and architectural discussions
• Evaluate solutions, e.g. via proof-of-concepts
• Design and execute benchmarks to evaluate performance, accuracy, and other NFRs.

Topics for Your Master’s Thesis

• GenAI copilot for automated dashboard generation
• or GenAI copilot for discovery of business process automation opportunities
• or GenAI tool to enrich business process models with information extracted from unstructured documents.

Expectations for applicant

• Feels comfortable working with Python and/or Java (preferably both)
• At least basic knowledge of Javascript/React code
• Knows how to write purposeful tests
• Ready to work with GenAI APIs and AI orchestration frameworks such as LangGraph
• Highly proactive in communication and ready to challenge requirements and think in terms of value delivered to users
• Always willing to collaborate with team members.

Supervisors

Why should you join us?

At Apromore, we're not just building software. We're revolutionizing how businesses tap into their data to unlock productivity and customer excellence.

We’re a global team of innovators collaborating across Australia, U.S., Germany, Finland, and Estonia to push the boundaries of process intelligence.

If you’re passionate about AI, data science, and high-end software products, and want to work in an environment where your ideas matter, Apromore is the place to be.

Application process

For the first round of admission, submit a link to a video of up to 10 minutes (shorter is OK) addressing the following points:

• Tell us about your software development skills and experience (open-ended question, anything you wish to tell us).

• How do you use copilots for software development? For which tasks do you see more value? Give us or show us concrete examples of how GenAI copilots have helped you and counter-examples where they have not been useful?

• In what other contexts have you used GenAI copilots (other than for software development)? Give us concrete examples of how you use them, including how do you usually prompt them and interact with them? What are they useful for? When have they been less useful to you?

• What skills and/or attitudes would you wish to develop during your industrial masters progam? Why do you think Apromore is a great place for you to develop these skills and/or attitudes?

If accepted to the second round, we will give you a small assignment consisting of developing a very small copilot using any LLM and API of your choice to answer certain types of questions (we will give you a spec with inputs and outputs your copilot should handle).

You will then be offered an interview where you will present to us your mini-copilot design and implementation, and we’ill ask you some technical and organizational questions (and you’ll also get a chance to ask us questions).

Company description

Introduction

More information

Location

The factory is located in Elva, close to Tartu.

Since the proposed topic is partly related to production equipment, then it requires the project participant to be present at the factory as well. All project contact persons have permanent positions in Elva.

Language

The working language of the project can be either Estonian or English.

Expected assignments

Mapping IoT implementation opportunities in production processes, in collaboration with the Elva production process engineers and production organization.

Development of a standard IoT platform that allows for improved traceability of process results in selected work sections.

Integrating different data structures to increase the efficiency of process performance management.

Topics for Your Master’s Thesis

Improving traceability of the production process using IoT solutions.

Development of a standardized data collection platform for a manufacturing company.

Expectations for applicant

• good teamwork skills;
• understanding of IoT operating principles and configuration;
• good level of programming skills;
• good knowledge of databases and APIs

Supervisors

Why should you join us?

Application process

Initial questions for the candidate, to which we expect a written answer:

1. What is your experience in developing manufacturing devices (including IoT devices) and platforms? Please describe some of the projects you have been involved in and what technologies you used.
2. How do you ensure the security and data privacy of IoT devices? Please provide examples of measures and best practices you have implemented.
3. What programming languages and tools do you think are most important in developing IoT solutions? How have you used them in your past projects?
4. How do you approach problem solving and troubleshooting in IoT systems? Please share an example of a situation where you had to solve a complex technical problem.

In the second round, we will interview candidates on site at the Elva factory.

About SensusQ

SensusQ transforms intelligence in the age of AI: how organisations turn vast, complex sets of data into actionable insights — in a fast and secure manner, with a human firmly on-the-loop.

For more information, visit the company website.

Introductory paragraph

In our dual-use platform, we blend decades of defence and intelligence experience with cutting-edge AI to build a platform that preserves data sovereignty, accelerates decision cycles, and empowers analysts and deciders alike to act with confidence.

By integrating multi-domain data sources, streamlining workflows, and enabling real-time access, we aim to make critical decisions more effective and reduce friction, not add it. What we build is more than software — it’s a force multiplier in the service of decisive, informed action, built from the ground up, for the age of AI.

Arrangement of collaboration

Location: Tallinn (hybrid is possible)

Format: In-person and remote mix (with the expectation of some on-site teamwork)

Language: English is our main working language.

Expected general assignments

• Automatically generate / update the knowledge graph, adding entities, relations, temporal links, and links to multimodal evidence.
• Integrate with the timeline & map components so that detected events can be visualised (map locations, time spans) and connected to KG entries.
• Work towards getting actionable insights by applying graph theory, as well as, graph neural networks based methods.
• Ensure the solution is scalable (works with large volumes of data, incremental updates, retraining/minibatching) and robust (noise, missing data, conflicting info).

Topic for master’s thesis

Graphusion - Leveraging Large Language Models in Order to Construct and Operationalize Knowledge Graphs

Take a look at the thesis proposal here.

(Technical) expectations for applying student

• Understanding of machine learning concepts and Large-Language Models
• Familiarity with Python and common data science libraries (NumPy, Pandas, TensorFlow or PyTorch)
• Familiarity with SQL
• Team player with a proactive approach to problem-solving
• Previous work with knowledge graphs is considered a merit

Supervisors

SensusQ: Jaan Übi, Lead Data Scientist

Institute of Computer Science: Eduard Barbu, Research Fellow in Natural Language Processing

Why should you join us?

Work in a startup on High-Impact, Real-World Problems - for the good of our clients, such as Eesti Laulu- ja Tantsupeo Sihtasutus, for the peace and security of Estonia and of Europe at large.

Data: our multimodel analytics platform is spatio-temporal, including a map component, a timeline, textual and image/video sources, a knowledge-graph - all of which is big data ready.

AI: in the form of AI, we have building-blocks, that tie all the data together and enrich it in a timely manner, with Agentic being the latest in our lego-set. For a data scientist in intelligence domain, data loves to live together - which is what model creators also aspire for, in order to create maximum value.

Great team atmosphere and company culture!

Application process

Round 1:

• Short video where you introduce yourself and share why you want to join us at SensusQ.
• What excites you about starting to work with knowledge graphs in a multimodal data environment?
• Describe a small project you’ve done and your biggest takeaway.
• How do you see this internship contributing to your long-term career goals?

Round 2:

• Virtual or in-person interview with our Data Science team. We’ll discuss your ideas and walk you through the day-to-day internship tasks.